Logs

I Tracked Every Meal for 90 Days. Here's What I Learned.

90 days of tracking every meal. Calories, macros, timing, mood, energy. The patterns I found surprised me. Most are not what nutrition advice would predict.

Eli Knight
By 3 min read
Share
On this page 11 sections
  1. 1 The setup
  2. 2 What the data showed
  3. 3 1. My calories varied way more than I thought
  4. 4 2. Protein varied even more
  5. 5 3. Meal timing didn't matter much for energy
  6. 6 4. Sugar before bed killed sleep quality
  7. 7 5. Workout days needed more food
  8. 8 6. Mood correlated with food less than I expected
  9. 9 What changed after the experiment
  10. 10 What surprised me most
  11. 11 This isn't advice
A food journal next to a healthy meal

For 90 days I logged every single thing I ate or drank. Calories, macros, timing, plus a one-line note about how I felt that day. Used a basic spreadsheet, not an app. Here's what the data showed.

The setup

What I tracked daily:

  • Every meal and snack — time and rough portions
  • Estimated calories and protein (carbs and fat estimated weekly)
  • Workout (yes/no and intensity)
  • Sleep duration
  • Energy 1-5
  • Mood 1-5
  • Body weight (morning, fasted)

I didn't change my diet during the experiment. Just tracked.

What the data showed

After 90 days, several patterns were clear.

1. My calories varied way more than I thought

I assumed my daily intake was stable. The data showed swings from 1,800 to 3,400 calories on different days. The average was around 2,500, but individual days varied dramatically.

The high-calorie days correlated with social meals and weekend evenings. The low-calorie days were workdays where I forgot to eat lunch.

2. Protein varied even more

I thought I was getting "enough" protein. The data showed 65g to 220g daily, average around 130g. The low days were lunch-skipped days. The high days were intentional workout days.

Below 100g, my next-day workout felt noticeably worse. Above 150g, no additional benefit I could measure.

3. Meal timing didn't matter much for energy

Conventional wisdom says eat breakfast for energy. My data showed no correlation between breakfast and afternoon energy ratings.

What did correlate with energy: total sleep duration the night before. Strong correlation. Much stronger than any food variable.

4. Sugar before bed killed sleep quality

Three desserts after 9 PM in the test period. All three nights showed worse sleep duration and worse next-day energy. Small sample size, but consistent.

5. Workout days needed more food

On training days, eating below 2,400 calories produced poor next-day energy. Above 2,400, no additional benefit. Rest days could go lower (around 2,000) without energy loss.

6. Mood correlated with food less than I expected

I assumed bad eating days would mean bad mood days. The correlation was weak. Mood correlated more strongly with sleep, exercise, and workload.

What changed after the experiment

I made three changes based on the data:

  1. Set a protein floor of 100g daily. Cheaper than tracking everything. Just hit the floor.
  2. Stopped eating sweets after 9 PM. Easy change with a clear benefit.
  3. Stopped worrying about meal timing. The data didn't support breakfast being critical for me.

I stopped tracking after the experiment ended. The system is too time-intensive for normal life. But the lessons stuck.

What surprised me most

How wrong my intuitions were about my own diet. I would have sworn my calories were stable and my protein was sufficient. The data showed neither was true.

Tracking for even a week or two would probably reveal patterns most people don't expect.

This isn't advice

This is one person's data. Your patterns will differ. The point isn't the specific findings — it's that tracking reveals things you can't guess.

If you're curious about your own patterns, try tracking for two weeks. You'll learn something.